Beyond the Noise: Effective Management of Non-Fatal Issues
summary
At Qonto Android teams handle about 600,000 non-fatal events/month across ~60 developers and 550 modules, and Crashlytics shows the events but Crashlytics lacks non-fatal velocity alerts. Teams enforce a contract to attach owner+priority metadata to each non-fatal, surface that metadata by rewriting the exception stacktrace or custom keys, then route alerts to Sentry where per-issue rules→Slack alerts notify the right channel — e.g., a FAQ webview 404 triggers an immediate Slack alert for the owning team.
AI-written summary · about
chapters (10)
- 0:00Non-fatal issues in banking apps
- 1:04Why non-fatal issues matter
- 3:53Challenges with 600K events
- 4:37Key questions for management
- 6:22Owner and priority in monitoring
- 8:50Firebase limitations for alerts
- 10:37Sentry fits non-fatal alerts
- 12:26Detecting missing FAQ articles
- 15:58Alerts shorten resolution chains
- 16:38Monitoring effectiveness takeaways
related talks
25mElevating Android App Resilience: Incident Management Strategies
44mEfficient Android app monitoring
35mMobile Observability at scale
37mYour Software is Broken — Pay Attention
19mYou Can't Stop Fires with an Ambulance
27mSensory Friendly Monitoring: Keeping the Noise Down
33mTackling Alert Fatigue with SLOs
19mHow to handle incidents at scale